Compendium of Evidence-Based Interventions and
Best Practices for HIV Prevention
Popular Opinion Leader (POL)
Men who frequent gay bars
Goals of Intervention
- Eliminate or reduce sex risk behaviors
Popular Opinion Leader (POL) is a community-level intervention designed to identify, enlist and train key opinion leaders to encourage safer sexual norms and behaviors within their social networks of friends and acquaintances through risk-reduction conversations. Cadres of trusted, well-liked men who frequent gay bars are trained to endorse safer sexual behaviors in casual, one-on-one conversations with peers at bars and other settings. During these conversations, the POL corrects misperceptions, discusses the importance of HIV prevention, describes strategies he uses to reduce his own risk (e.g., keeping condoms nearby, avoiding sex when intoxicated, resisting coercion for unsafe sex), and recommends that the peer adopt safer sex behaviors. Popular opinion leaders wear buttons displaying the project logo, which also is on posters around the bars, as a conversation-starting technique. Each POL agrees to have at least 14 conversations with peers and to recruit another POL. POLs attend four weekly 90-minute training sessions that involve didactic and group discussions, modeling of effective health promotion messages, and extensive role play.
- Diffusion of Innovations
Popular opinion leaders who are trusted, well-liked men frequenting gay bars, and trained to endorse safer sex behaviors
- POL training
- Goal setting
- Group discussion
- Role play
- Skill building
- Casual, one-on-one conversations
- Printed materials (logos, symbols, other devices)
An intervention package was developed with funding from CDC’s Replicating Effective Programs (REP) Project. The intervention package and training are available through CDC’s Diffusion of Effective Behavioral Interventions (DEBI) project.
The original evaluation was conducted in Biloxi, MS; Hattiesburg, MS; and Monroe, LA between 1989 and 1991.
Key Intervention Effects
- Reduced unprotected anal intercourse
- Increased condom use
- Reduced number of sex partners
The baseline study sample of 659 men in 3 communities is characterized by the following:
- 86% White; 14% African American or Hispanic
- 100% Male
- 100% MSM
- Mean age of 29 years
- Mean education level:15 years
- The eligible communities were small cities (50,000 to 75,000 residents) that were separated by at least 60 miles from another city included in study or any other larger city (> 75,000 residents) and had one or two large gay bars.
- Men were eligible for assessment if they frequented gay bars in the eligible communities.
One city (Biloxi, Mississippi) was randomly selected to receive the intervention and the remaining two cities (Hattiesburg and Monroe, Mississippi) served as comparison cities.
The comparison cities received no specific intervention, but AIDS prevention brochures and posters were generally available in gay bars.
Relevant Outcomes Measured and Follow-up Time
- Sex behaviors during past 2 months (including any unprotected anal intercourse, unprotected receptive anal intercourse, or unprotected insertive anal intercourse; condom use during any anal intercourse, receptive anal intercourse, or insertive anal intercourse; and number of sex partners) were measured at baseline and 3 to 6 months post intervention
Not applicable due to cross-sectional samples1
- Participants in the intervention city reported significantly greater reductions in any unprotected anal intercourse (including any, receptive or insertive) than participants in the two comparison cities (all p’s < .05).
- Participants in the intervention city reported significantly greater increases in condom use during anal intercourse (including any, receptive or insertive) than participants in the comparison cities (all p’s < .05).
- Participants in the intervention city reported significantly greater reductions in having multiple sex partners than participants in control communities (p < .05).
- This intervention fails to meet the best-evidence criteria due to insufficient number of communities (1 city received the intervention; 2 cities served as comparison cities).
- Intervention efficacy was evaluated by comparing the first city to receive the intervention (Biloxi) to two comparison cities (Hattiesburg and Monroe). The intervention was evaluated in all three cities using a sequential stepwise lagged design. The intervention produced reductions in high-risk behaviors (e.g., unprotected anal intercourse) with the same pattern of effects replicated in all three cities (Kelly et al., 1992).
- Additional evidence is found in a replication study conducted by Kelly et al., 1997. However, this replication intervention fails to meet the GOOD-EVIDENCE criteria due to excluding 8 (4 intervention and 4 comparison) of 16 intended communities and dropping participants who were transient or exclusively partnered (32% of baseline assessment and 39% of follow-up assessment) from the analysis.
- Additional evidence is found in an adaptation study conducted by Jones et al., 2008 for African American MSM. This adaptation study also fails to meet the GOOD-EVIDENCE criteria due to lack of a comparison arm.
- Kelly, J. A. et al. (1991). HIV risk behavior reduction following intervention with key opinion leaders of population: An experimental analysis. American Journal of Public Health, 81, 168-171.
- Kelly, J. A. et al. (1992). Community AIDS/HIV risk reduction: The effects of endorsements by popular people in three cities. American Journal of Public Health, 82, 1483-1489.
- Kelly, J. et al. (1997). Randomised, controlled, community-level intervention, HIV-prevention intervention for sexual risk behaviour among homosexual men in US cities. The Lancet, 350, 1500-5.
- Jones, K. et al. (2008). Evaluation of an HIV prevention intervention adapted for Black men who have sex with men. American Journal of Public Health, 98, 1043-1050.
Jeffrey A. Kelly, PhD
Department of Psychiatry and Mental Health Sciences
Medical College of Wisconsin
701 Watertown Plank Road
Milwaukee, WI 53226